State job statistics underestimated losses

The Associated Press
Wednesday March 27, 2002

SACRAMENTO— Government statistics appear to have underestimated job losses last year in California, a sign that the recession was worse than previously thought, according to a newspaper report Tuesday. 

New data shows that the state had about 180,000 fewer jobs in September than officially reported by the federal Bureau of Labor Statistics and the state Employment Development Department, according to confidential data obtained by The San Francisco Chronicle. 

The new statistics, from unemployment insurance tax filings, normally are made public only once a year. 

The miscount raises the question of whether official employment statistics gave an accurate portrait of the economy and whether that data can be relied on by planners. 

For example, 180,000 fewer jobholders would mean an unanticipated loss of roughly $200 million in annual tax revenue for California, said Ted Gibson, former chief economist for the state Finance Department. 

That’s something the state can ill afford when it faces a budget deficit of as much as $15 billion. 

The shortfall represented about 1.2 percent of the state’s total of 14.7 million reported jobs in September, the most recent month for which new data are available. 

Bureau officials said they had begun a review to find out why the new data showed so many fewer jobs than official employment statistics. 

“We do see a bigger difference than we are comfortable with,” said Pat Getz, the bureau’s division chief for current employment statistics. 

The agency couldn’t yet explain the divergence, Getz said. 

State Employment Department officials confirmed the accuracy of the 180,000 September job shortfall in California but played down its significance. Different employment data sources are “never in complete agreement,” said Richard Holden, chief of the department’s labor market information division. 

At issue is the so-called nonfarm payroll jobs numbers released monthly by the bureau for the nation and each of the 50 states. They rank second only to the official unemployment rate in the attention they get as measures of the labor market’s health. 

The payroll numbers are based on a survey of employers designed by the bureau and conducted by state labor departments. In California, the Employment Department collects data from just under 40,000 establishments across the full range of industries. 

But the payroll numbers are estimates rather than strict head counts and are subject to errors inherent in deriving statistics from a sample. 

For that reason, the bureau supplements its monthly survey by tracking a separate set of data — quarterly unemployment insurance tax filings. Those filings by employers provide the closest thing to an actual count of jobs. 

Once a year, the bureau releases a formal revision of the less accurate payroll jobs figures based on information from these filings. 

“That is the bible,” said James Glassman, senior U.S. economist with J.P. Morgan Securities about the quarterly tax filing data. “It is viewed as the more accurate view of the labor market.” 

The gap between the number of California jobs found in the payroll survey and the number counted in quarterly tax filings has widened dramatically.